Innovation studies have been conducted in various fields of research during the time (Baregheh, Rowley & Sambrook, 2009). Such studies have been associated with productivity, growth, business development and similar concepts (Fagerberg, 2005; Kaplinsky et al., 2009, Khajeheian, 2014). Innovation promotes entrepreneurship and new venture creation (Khajeheian, 2013; Salamzadeh and Kawamorita, 2017) and it allows companies to stay competitive (Khajeheian, 2016). Considering such impact, innovation has always been a popular research topic for many of researchers in the field of management as well as practitioners in the field of Information Systems (Avgerou, 2008). Gonzalez-Perez and Velez-Ocampo argues that managerial commitment lowers the cost of adaptation with technology (2014, p. 541). However, the increasing application of innovations, especially technological innovations in services, is not without challenge and part of the challenge come from the acceptance of Innovation and technology readiness. By Parasuraman & Colby (2015), Technology readiness describes the individual’s propensity to use new technologies for different purposes. Technology readiness associates with Innovation acceptance (Lippert and Govindarajulu, 2015) and is one of the important factors that determine the success of application of an innovation in a society.

Parasuraman (2000) developed a 36-item technology readiness index (TRI). That scale consisting of four dimensions: optimism, innovativeness, insecurity and discomfort. Since then, technology readiness index has been applied for example in the context of self-service technologies (Gelderman et al.2011; Liljander et al., 2006), mobile services (Chen et al., 2013; Sophonthummapharn & Tesar, 2007) and wireless technology users (Chang & Kannan, 2006). Researches have incorporated technology readiness with other technology adoption models such as Unified Theory of Acceptance and Use of Technology (UTAUT) (Chiu, Fang, & Tseng, 2010), Technology Acceptance Model (TAM) (Lin et al.2007; Walczuch et al., 2007) and Expectation-Confirmation Model (Chen et al., 2013). Also cross-cultural validity of the scale has been assessed (Meng et al., 2009). Technologies change over time and the rapid development of new technologies led to the development of an updated and streamlined technology readiness index, that called TRI 2.0 (Parasuraman & Colby, 2015). TRI 2.0 scale has been updated to match with the recent changes in the technology environment, and at the same time the scale has been streamlined to a compact 16- item scale to be more easily adopted as a part of research questionnaires.

the Technology Acceptance Model (TAM) of Davis et al (Davis,1989; Davis, Bagozzi, Warshaw,1989) is one of the most widely used models of information systems (IS) that explain or predict the motivational factors in user acceptance of technology. Perceived ease of use refers to the degree to which an individual belief that the use of an information system is free from effort, while perceived usefulness is a perception that using a particular system enhances one’s job performance (Davis et al., 1989). It is generally assumed that personality does influence an individual’s technology adoption behavior (McElroy et al., 2007; Devaraj et al., 2008). Thus technology acceptance model and technology readiness have been integrated in the prior research (e.g. Walczuch et al., 2007; Lin et al., 2007).

Igbaria and Tan (1997) presents relationship of IT acceptance and individual impact, that was based on the theoretical perspectives of DeLone and McLean and conceptualization of Rogers (1983) of innovation. As it is shown in Figure 1, this model includes three variables of user satisfaction, system usage and individual impact. This model proposed that the degree of user satisfaction has a positive effect on the level of usage. Also user satisfaction and system usage effect on individual impact.

Source: Igbaria and Tan, 1997, p. 115.

Two variables of user satisfaction and system usage are frequently used for evaluation of IT success. Fishbein and Ajzen (1975) state that individual’s attitudes play a vital role in the respective behavior. Some other researchers confirm the relation of user attitudes to their technology acceptance (Dillon and Morris, 1996; Venkatesh et al., 2003; Mun et al., 2006; Bhatti, 2007).

The American Customer Satisfaction Index (ACSI) raised based on the Swedish Customer Satisfaction Barometer (SCSB), that successfully introduced in 1989 as a tool to assess the customer satisfaction (Fomell, 1992). The ACSI based on two theories of QSP (Quality, satisfaction and performance) and Hirschman’s exit-voice theory (1970). This model measures the cause and effect relationship that drivers of satisfactions (customer expectations, perceived service quality, perceived value) are on the left side, satisfaction in the center and outcomes of satisfaction (customer complaints, customer loyalty) on the right side. (see Figure 2).

Model of Information Systems Success of DeLone and McLean (2003) incorporates six elements of system quality, information quality, use, user satisfaction, individual impact, and organizational impact. Information quality (semantic level) and system quality (technical levels) are two key antecedents of user satisfaction. Information quality assessed by measuring information attributes and system quality assessed by measuring of ease of use (Doll & Torkzadeh, 1988; Rai & Welker, 2002). DeLone and McLean added service quality to the original model and identified it as an additional antecedent of user satisfaction. Figure 3 shows their model.

Factors affecting this model are defined as follows:

• Quality System: Adaptability, Availability, Reliability, Response Time, Usability

• Quality of Information: Completeness, Ease of Understanding, Personalization, Relevance, Security

• Service Quality: Assurance, Empathy, Responsiveness

• Use: Nature of Use, Number of Visits to the Sites, Number of On-Line Transactions

• User Satisfaction: Repeated Purchase, Repeated On-Line Purchase, Applied Survey

• Advantages: Cost Savings, Expansion of Markets, Additional Sales, Reduced Search Cost, Time Saving. (DeLone-McLean, 2003)

Three studies have been conducted regarding the evaluation of satisfaction rate of users of information systems in Azerbaijan, Malaysia and Taiwan.

Government of the Republic of Azerbaijan has undertaken a series of corrective actions to be linked with the global economy and enhance the quality of people’s lives including changes in the tax system that began in 2005. Using the Automated Tax Information System (ATIS) in 2006 could be considered as one of the achievements that led to the creation of an e-government initiative. Its main purpose is to provide an opportunity for taxpayers to fill out and submit their income tax returns via the Internet without having to go to the tax oftce. In the country’s bid to the modernize the government, VAT and electronic invoices were introduced as a product since January 2010. As a result, taxpayers can send electronic invoices for sold products and provided services. For Satisfaction Evaluation, the first step should be providing services to assess knowledge and satisfaction. In the next stage, taxpayers should be aware of the existence of this service. The third step should be taken to evaluate the knowledge and satisfaction. And the final step is to improve the system with regard with the results of previous step. In the survey implemented in August-September 2010, Key parts of the final report on behalf of the selected areas were determined to check every service structure as follows: General Information Services, the level of understanding of the services, the differentiation of services and the level of public satisfaction, Recommendations for improving service. (Shikhaliyev,2011)

In 2006, the Malaysian Inland Revenue Board (MRIB)proposed the adoption of electronic forms as part of an e-government initiative. This was to happen in the context of customer to government relationship., Here, the users could fill out forms manually or electronically. Since then, electronic forms have been on the increase in Malaysia. However, the issues related to the quality of electronic services lead to user dissatisfaction. Although the number of e-forms has increased, the issue of user dissatisfaction discussed above impeded the achievement of the objectives for adopting electronic forms by up to 80%. Hence, considering the nature of the system quality is vital. This research is based on an updated version of the Delone- McLean model. Here the three factors of service quality, information quality and system quality have been considered in order to achieve information systems success, where the following results have been achieved: 1) the data quality has a direct relationship with the users’ consent, 2) the existence of a quality system is directly related to user satisfaction, 3) Quality of service is directly related to the user satisfaction, 4) User satisfaction is directly related to the main perceived benefits (convenience) of users, 5) User satisfaction has a direct relationship to the main perceived benefits (eftciency) of users (Islam et al., 2012).

In Taiwan, the National Tax Administration (NTA) is responsible for collecting tax returns and the income tax returns are filled out in three ways namely: manually, electronically and via the usage of two-dimensional barcodes. Although, the use of barcodes is considered as an electronic method. Taiwan launched its on-line tax-filing system since 1998. The on-line tax-filing system reduced the costs of printing and distribution. The satisfaction of the users of the system is based on the usefulness of the system, in which, DeLone- McLean model is used. In this study, the information quality, system and service quality have been divided into the sub- factors as follow:

1) Information Quality: a) Informative, b) Accuracy; 2) System Quality: a) Access, b) Interactivity, c) Ease of use; 3) Service Quality: a) Responsiveness, b) Reliability, c) Empathy (Chen, 2010).

The sample population of this research was selected from value added taxpayers in Tehran. The random sampling method was used for sampling. The five scale Likert questionnaire were distributed among the taxpayers at the Tehran Taxpaying services for 15 days. Since, the VAT system has greatly reduced the number of in-person refers, and many real taxpayers use internet café services, the number of respondents produced 150 questionnaires. After removing wrongly filled questionnaires, 102 questionnaires formed the sample of study. The questionnaire is attached in the appendix of the paper.

ACSI is the basis for this model. The variables: taxpayer exceptions, perceived quality, perceived value and satisfaction are extracted from this model with the difference that loyalty is not applicable for taxpayers, because there is no other alternative to compete with. Since perceived quality in the ACSI has an effect on Satisfaction and based on DeLone -McLean model information service and system quality is related to satisfaction, so we can extend the perceived quality on ASCI model to service quality, system quality and information quality. In other words, each of the three sections do have an impact on the perceived quality from user and consequently on its satisfaction.

Increasing awareness and knowledge is one of factors that was studied in Azerbaijan which has an effect on satisfaction and the acceptance system by users.

In this model, for increasing awareness, training was considered and the training has been named as one of the expectations from taxpayer. Especially, given that many users are advanced in age and are phrase not computer literate. Another user expectation is to respect for their rights. In some cases, may be due to system errors such as disruption of the Internet, the user wiped out part of the information. In this case, appropriate solutions must exist for respecting the rights of the users. On the other hand, the integrated tax systems; especially in the audit section, creates the need to establish a uniform procedure for the whole country. This will also prevent discrimination between taxpayers.

In the complaints section, attention has been drawn to two issues. The first complain is as result of the establishment of procedures and legal processes and may indirectly lead to dissatisfaction with the system. The second complain is specific to the system, such as system outages, sluggishness and the lack of confidence in the system which directly leads to the user dissatisfaction. Thus, the framework of this research is presented based on the following: expectations and perceived quality effect on the value perceived; the value perceived influences on the user satisfaction and that increased satisfaction will have an influence on the rate of compliant effectively.

Eleven hypotheses were extracted to be tested:

• The user complaints on the rules and procedures has an impact on his/her dissatisfaction of the system.

• Facilitating and the quality of service provided by the system has an impact on user satisfaction of the system.

• Proper training has an impact on user satisfaction of the system.

• Respecting the rights of user has an impact on his/her satisfaction of the system.

• Information quality has an impact on user satisfaction of the system.

• Service quality has an impact on user satisfaction of the system.

• System quality has an impact on user satisfaction of the system.

• Meeting the perceived quality has an impact on user satisfaction of the system.

• Meeting the user expectations of the value-added tax system has an impact on user satisfaction of the system.

• Perceived value has an impact on user satisfaction of the system.

• Reduction of complaints increases user satisfaction of the system.

SPSS software was used to analyze the data. Descriptive statistics involving central indicators and dispersion in the form of mean and standard deviation on one hand, and inferential statistics in the form of calculations of correlation coefficients and linear regression model on the other hand. Then, the standardized coefficient beta (β) has been used in the regression equations to determine the presence and power of corresponds in a given model.

The following information is obtained of all the respondents:

Table 3 shows the mean and standard deviations values for independent variables.

According to Table 4, it can be seen that the mean of the variable, user complaints of rules and procedures is 2.87; the mean of the variable, facilitation and expedition of the affairs by the system is 3.42; the mean of the variable, proper training is 3.08; the mean the variable, respecting the user rights is 3.65; the mean of the variable, the information quality is 3.14, the mean of the variable , service quality is 3.22; the mean of the variable, system quality is 3.35; the mean of the variable, meeting the perceived quality is 3.10; the mean of the variable, meeting user expectations is 3.52; the mean of the variable, perceived value is 3.09, and the mean of the variable, complaint decrease is 3.48.

In the table below, the correlation coefficient between independent and dependent variables of the study is examined.

According to the table, there is a relationship at the significance level of 0.05 between the independent variables and dependent variables (except user complaints of rules and procedures that has the lowest correlation with the degree of user satisfactions; so, it was not significant) in all cases. Among the independent variables, the variable that indicates the meeting of the perceived quality with the correlation coefficient of 0.744 has the highest correlation with respect to user satisfaction. Likewise, the variable that indicates proper training has the lowest correlation with respect to user satisfaction.

As it was mentioned earlier, the linear regression model has been used to test the hypotheses. The level of significance of regression that is shown with beta standardized coefficient is considered as the basis of inference to display existence and the power of relations between variables in all the assumptions. Table 4 shows the results of hypotheses tests in the form of regression model.

The result of the first hypothesis: The modified coefficient of determination (R2) is 0.4% of the available variance in the user satisfaction variable. The standardized coefficient beta is not significant at 0.05 level for the variable of user complaints of the rules and procedures (β= -0.063). Accordingly, we reject the first hypothesis. The results show that Durbin-Watson statistic is ranged from 1.5 to 2.5 and there is no strong autocorrelation between the errors of the regression model. The lack of autocorrelation between errors is accepted as one of the basic assumptions of the regression in the fitted model.

The result of the second hypothesis: As it is seen, the modified coefficient of determination (R2) explains 23.4% of the available variance in user satisfaction variable. The standardized coefficient beta is significant at 0.05 level for facilitating and expediting the affairs by the system’s variable (β= 0.484). Accordingly, the second hypothesis will be accepted. The results show that Durbin-Watson statistic is ranged from 1.5 to 2.5. There is no strong autocorrelation between the errors of the regression model.

The lack of autocorrelation between errors is accepted as one of the basic assumptions of the regression in the fitted model.

The result of the third hypothesis: The modified coefficient of determination (R2) explains 4.3% of the available variance in user satisfaction variable. The standardized coefficient beta is significant at 0.05 level for variable of proper training (β = 0.207). Accordingly, the third hypothesis will be accepted. The results show that Durbin- Watson statistic is ranged from 1.5 to 2.5. There is no strong autocorrelation between the errors of the regression model. The lack of autocorrelation between errors is accepted as one of the basic assumptions of the regression in the fitted model.

The result of the fourth hypothesis: The modified coefficient of determination (R2) explains 17.4% of the available variance in user satisfaction variable. The standardized coefficient beta is significant at 0.05 level for variable of respecting user rights (β= 0.417). Accordingly, the fourth hypothesis will be accepted. The results show that Durbin-Watson statistic is ranged from 1.5 to 2.5. There is no strong autocorrelation between the errors of the regression model. The lack of autocorrelation between errors is accepted as one of the basic assumptions of the regression in the fitted model.

The result of the fifth hypothesis: The modified coefficient of determination (R2) explains 41.5% of the available variance in user satisfaction variable. The standardized coefficient beta is significant at 0.05 level for information quality variable (β = 0.644). Accordingly,

the fifth hypothesis will be accepted. The results show that Durbin- Watson statistic is ranged from 1.5 to 2.5. There is no strong autocorrelation between the errors of the regression model. The lack of autocorrelation between errors is accepted as one of the basic assumptions of the regression in the fitted model.

The result of the sixth hypothesis: The modified coefficient of determination (R2) explains 28.9% of the available variance in user satisfaction variable. The standardized coefficient beta is significant at 0.05 level for service quality variable (β= 0.537). Accordingly, the sixth hypothesis will be accepted. The results show that Durbin- Watson statistic is ranged from 1.5 to 2.5. There is no strong autocorrelation between the errors of the regression model. The lack of autocorrelation between errors is accepted as one of the basic assumptions of the regression in the fitted model.

The result of the seventh hypothesis: The modified coefficient of determination (R2) explains 22.3% of the available variance in user satisfaction variable. The standardized coefficient beta is significant at 0.05 level for system quality variable (β= 0.472). Accordingly, the seventh hypothesis will be accepted. The results show that Durbin- Watson statistic is ranged from 1.5 to 2.5. There is no strong autocorrelation between the errors of the regression model. The lack of autocorrelation between errors is accepted as one of the basic assumptions of the regression in the fitted model.

The result of the eighth hypothesis: The modified coefficient of determination (R2) explains 55.3% of the available variance in user satisfaction variable. The standardized coefficient beta) is significant at 0.05 level for meeting perceived quality variable (β= 0.744). Accordingly, the eighth hypothesis will be accepted. The results show that Durbin-Watson statistic is ranged from 1.5 to 2.5. There is no strong autocorrelation between the errors of the regression model. The lack of autocorrelation between errors is accepted as one of the basic assumptions of the regression in the fitted model.

The result of the ninth hypothesis: The modified coefficient of determination (R2) explains 27.6% of the available variance in user satisfaction variable. The standardized coefficient beta is significant at 0.05 level for meeting user expectations variable (β= 0.525). Accordingly, the ninth hypothesis will be accepted. The results show that Durbin-Watson statistic is ranged from 1.5 to 2.5. There is no strong autocorrelation between the errors of the regression model. The lack of autocorrelation between errors is accepted as one of the basic assumptions of the regression in the fitted model.

The result of the tenth hypothesis: The modified coefficient of determination (R2) explains 21% of the available variance in user satisfaction variable. The standardized coefficient beta is significant at 0.05 level for perceived value variable (β= 0.458). Accordingly, the tenth hypothesis will be accepted. The results show that Durbin- Watson statistic is ranged from 1.5 to 2.5. There is no strong autocorrelation between the errors of the regression model. The lack of autocorrelation between errors is accepted as one of the basic assumptions of the regression in the fitted model.

The result of the eleventh hypothesis: The modified coefficient of determination (R2) explains 31.2% of the available variance in user satisfaction variable. The standardized coefficient beta is significant at 0.05 level for complaint decrease variable (β= 0.559). Accordingly, the eleventh hypothesis will be accepted. The results show that Durbin- Watson statistic is ranged from 1.5 to 2.5. There is no strong autocorrelation between the errors of the regression model. The lack of autocorrelation between errors is accepted as one of the basic assumptions of the regression in the fitted model.

Totally, 10 hypotheses out of 11 research hypotheses are confirmed and one hypothesis is not confirmed. Also, satisfaction of the value added tax system was assessed at 63.79%.